Krea adds Google's Nano Banana 2 image model
Krea says Google DeepMind's Nano Banana 2 is now on its platform, promising faster 4K generation with better text rendering and character consistency, though Google itself has not supplied benchmarks or pricing.
Creative platform Krea announced that Nano Banana 2 (NB2), which it describes as the next generation of Google DeepMind's image model, is available on its service now. Krea pitches it as "Pro-quality images at flash speed."
According to Krea, NB2 can:
- render readable text inside images in multiple languages, for uses such as posters, social cards, UI mockups and infographics;
- keep up to five characters and fourteen distinct objects consistent within a single scene, and keep characters recognizable across a series of images;
- build structured visuals such as diagrams and data visualizations, using what Krea calls "world knowledge from Gemini";
- follow long, detailed prompts that specify lighting, camera angles and material textures without dropping parts of the instruction.
Output resolution runs from 512px to 4K across a range of aspect ratios. Krea also says lighting, textures, fine detail and composition are better than before, but it gives no metrics to back that up.
The post is Krea's own product copy. It includes no benchmark results, pricing, rate limits or technical details, and it does not quote Google or link to any Google DeepMind material. It does not say whether NB2 is available through Google's own products or APIs, or how it compares with earlier Nano Banana models beyond the general claims of improvement. Until Google publishes its own documentation, the performance claims cannot be checked independently.